Wednesday, June 22, 2011

Understanding the Grid

I was reading through Recivilization's theory on the three separate sources of American culture earlier, and how these three sources manifested themselves in urban design. As the writer notes, the Pennsylvanian pattern--the grid--came to win out over New England's medieval mores and the South's Baroqueness--but, unfortunately, he offers no in-depth analysis of the Pennsylvania tradition.
A boring grid.
This is unfortunate. While the gird easily became a curmudgeonly speculators' tool, Pennsylvanians have, in fact, managed to make it just as malleable and sophisticated as both the medieval and baroque traditions. This is especially noticeable in Philadelphia, where the grid was developed and later perfected--it is unfortunate that only one variant of this whole system of urban design came to dominate every other American city. Old Urbanists, like Old Urbanist, Recivilization, and Nathan Lewis, have a tendency to look on grids with the same scorn they attribute to the Modernists' looking on Trumbauer and Beaux-Arts urban design,* but like it or not, a grid's what we got, so it's far better working on how to perfect it rather than wax nostalgic about gridlessness. How else are we supposed to actually fix broken cities?

How was this done? Three ways: the use of the grid as a spatial organizer at the neighborhood level, and secondly by the use of main streets to reference and reinforce the grid, linking it between neighborhoods; the third element--perfected in Philadelphia's  urban paradigm but seemingly lost outside of it--is the use of sub-grids, with their own internal spatial organization and reference-and-reinforcement networks, is connected with art of the grid in overarching city design, and seems a necessity to retain the grid's usefulness in terms of spatial organization without impairing it by overextension.
Internal Spatial Organization in Wash West. Note mews.
1. Internal spatial organization. The grid divides blocks, but does not allocate within blocks. These blocks are then subdivided, with new streets run through. In the older Colonial and antebellum days of the city, the pattern was one of conch-shell mews; during the Victorian era, the pattern became the use of one or two through-running minor streets. These streets normally run only one or two blocks, and thus have terminated vistas at both ends, giving them an aura of a public room. Even in the poorest parts of North Philadelphia, the stronger individual blocks are usually (though not without exception) planted along one of these minor streets. The effect is, again, one of mews.
The blue lines reference and reinforce the grid; the red lines break different grid alignments; the brown lines transect the grid without breaking it.
2. Reference and reinforcement. One of the more remarkable things about Philadelphia's grid is how it references and reinforces itself. Unlike the City Commissioners' plan for New York, there does not seem to be a clearly defined plan for Philadelphia. Indeed, neighborhoods such as Northern Liberties and Southwark were not actually in Philadelphia city as late as 1840--and these communities retain a quite strong connection with the city's internal grid.

How did this come to be? Reference and reinforcement. A reference is a main street a grid branches from; as the grid extends further from this street, another main street--a reinforcement--continues it. The part of West Philadelphia dominated by Lancaster Avenue is a perfect example of reference and reinforcement. Since Lancaster is the principal main street in the area, shouldn't it be a local grid reference? No...actually, the grid Lancaster exists in--and cuts across--references Powelton Avenue, and is later reinforced along 42nd St., thence Girard Avenue, and finally 63rd. Similarly, Southwest Philly's primary grid references Springfield Avenue first, and the Woodland Avenue (west of 49th), and is reinforced by Whitby and Elmwood Avenues.

The final element of reference and reinforcement is grid-breaking streets (or foils). These take two forms: either a street will transect an existing grid without getting referenced or reinforcing (such as Frankford, Ridge, Lancaster, or Passyunk), or it will divide different grid sections (such as Haverford, Baltimore, or Kensington). Both are necessary, as they interject variety into the otherwise monotonous grid.

Without reference and reinforcement, a grid will unravel itself. Market Street, once it becomes West Chester Pike, ceases to be referenced; the overarching grid unravels and becomes a series of highly localized mini-grids.
Multiple grids in Philadelphia
3. Multiple grids. Chicago is completely dominated by a singular grid. Many American cities, such as Houston and Minneapolis, have a downtown grid set at a 45-degree bevel to the overarching urban residential grid. There are only one or two grids in most major American cities.

San Francisco is an exception to this rule, and in fact its main street (Market Street), functions as a foil, effecting the 45-degree split between the downtown and circumferential grids right in the middle of downtown. In addition, San Francisco's circumferential grid has some occasional kinks in it--a slight change of orientation following the reference and reinforcement streets, but no new references or reinforcements.

In Philadelphia, however, the situation is different. There is no circumferential grid, and only about half of the city is defined by the primary grid--referencing Market and Front, and reinforced primarily by Broad, 25th, and 30th. This grid looks like a north-south column with a long westerly branch. The rest of the city is defined by a succession of lesser grids: Southwest, referencing Springfield-49th-Woodland, and reinforced by Elmwood and Whitby; the River Wards, referencing Girard at first and then Richmond, and reinforced by Emerald; the Lower Northeast, referencing Cottman and reinforced by Devereux, Pratt, and Castor; and Powelton, referencing Powelton and reinforced by 42nd, Girard, and 63rd.
Manayunk and Roxborough: looks like a grid, but really isn't.
But this only accounts for so much. In fact, the most surprising thing about Northwest Philadelphia is that it's not quite gridded so much as grid-like. Looking from afar, and walking around, many Northwest Philadelphia neighborhoods feels like walking in a grid, but when you look closer, there's no reference-and-reinforcement system in place, and actual alignments shift almost at will. There is no real grid in lower East Falls, for example; only an illusion of one created by density between Calumet and Indian Queen. Similarly, Manayunk and Roxborough streets act like a grid at first but shifts in alignment may or may not be taken account, and neither Main nor Ridge are strong references. Grid-like systems describe the urban setup of many medieval-era European cities, which likewise organically expanded in a similarly grid-like pattern--which, again, indirectly points out the history of the city.

(Edit) I should talk about squares some. Squares are key public amenities that occur in any city building scheme. One of the great failures of the American grid is the failure to provide any. Even here, Victorian Philadelphia fails: the most important public squares, such as Rittenhouse or Franklin or Passyunk or Liberty Lands--and so on--either predate or postdate the Victorian era. Even today, carving out public squares in gentrifying neighborhoods is a strenuous issue. As such, every grid, to succeed, also has to have an internal method of providing public squares. The Billy Penn grid was one-off; Savannah's grid actually works best in this regard.

In sum, however, there are urban design principles in play in grids which do allow them, when those techniques are used properly, to grow organically and create a satisfying sense of place. Secondly, when the techniques are not present but an urban place where it looks like they ought to be is, this place is grid-like and in all likelihood evolved organically in the same way Koln's city center evolved organically out of the Roman grid of Colonia.
* Scientistic urban planning was led, in large part, by Lewis Mumford, who, like Carmello Sitte, idealized medieval urban design. Mumford was clearly influenced by Sitte (do read The Culture of Cities), which makes it kind of odd (and ironic) that other Modernist urban planners--especially those working under the Radiant City frame of reference--who would not so easily dismiss Mumford dismissed Sitte based on no more than some poor translation choices.


  1. What's not to like about grids?

  2. Steve -- my objection isn't to the grid itself, but to inflexibility and absolute uniformity in any plan, whether orthogonal, radial or any other type.

    If the grid does not require those outcomes, it does at least encourage them. A grid has no natural end point, and few gridded cities really contemplated interior block space being subdivided by streets beyond perhaps a single central alleyway. Although a thoughtful planner can produce an effective irregular grid (I've seen good examples from New Urbanist developments), it's far easier to simply default to mindless repetition of standard-sized blocks on standard-width streets, ad infinitum. This is, I think, why so many "bad" grids were produced in 19th century America -- it wasn't simply misfortune.

    One possible alternative is to imagine a type of planning that is neither all-encompassing nor totally hands-off, which plans certain routes but intentionally leaves room for others to fill in through incremental growth. From what I've read, a similar system guided the development of New York through the mid-18th century, and probably many other cities of that era. It would, however, involve abandonment of the grid so far as "grid" is taken to represent the idea that the planner ought to be responsible for laying out all streets in a town.

  3. Charlie: I wasn't disputing this--what I was taking issue with is the fact that Recivilization has chapters on how to do it right in the medieval (New England) and Baroque (Southern) styles...but not with a grid. Philadelphia manages to get it right, mainly because the irregular streets predate the grid's existence, but many other cities fail to.

    My thesis is that grids can produce satisfying results, if the right methods are used, just as any other method of city-building can. But to do so, they can't be the be-all-end-all, and in fact, they need to have areas where the grid changes or simply breaks down. Monotony is the problem.

  4. In entry 111 on Recivilization ("the american grid"), there is at least a hint of an answer this question about the grid:

    "In all the older towns, commercial pressure cheapened the Baroque or medieval plans left from the original colonies. Throughout history, from Hippodamus to William Penn, grids had been only the background to sophisticated yet simple design systems, where amenity and practicality were held in a satisfying balance. Now, in the New Secular Order, continuity with the past was forsaken; shorn of its embellishments, its open spaces, boulevards and focal points, the grid became a radically new type of plan, one devoted to no finer purpose than facilitating speculation in city land, and religiously excluding anything that might hinder speculation."

    So maybe for the author the Baroque refinement of the basic rectangular plan does represent the ideal form of the grid -- it is simply a grid that takes account of amenity. Boulevards and focal points break the monotony. At least that's the way I am reading it, right or wrong.

  5. Perhaps. In a way, L'Enfant's Washington may be the most well-designed grid, since the whole construct is done in such a way that the major thoroughfares cut through the primary grid, rather than structure it. But the sheer prevalence of such thoroughfares--far more than the more naturally-evolved Philadelphia grid--is certainly a Baroque excess.

    If I were designing a new city, I would basically have the major streets be the pikes that already serve the area and other important streets be laid out along existing farmland boundaries. Then the next step would be to turn the farm lanes (all farms have internal access lanes) into neighborhood thoroughfares, and finally branch the street and lot network off that. In this paradigm, the pike would be a 60-80 ft. complete street, the wider boundary roads and converted access roads about 30 ft., complete or sharrowed, and the neighborhood streets would be 15-20 ft. naked streets. Approximately 70% of the network would be 20 ft. or less, 20% 30-50 ft., and 10% 60 ft. or more.

    Alternatively, without the starting point farms offer, a sophisticated grid can have the same function, as long as the complete streets are far enough apart, the subdivisions the same size as traditional farms (~40 acres), and the internal street network again using a highly-interconnected naked streets (mews) network transected by wider thoroughfares. The ruling grid in Phoenix would be a perfect starting-out point for such a mode of design--but instead the subdivisions there were all filled with, well, subdivisions. Such opportunity wasted.

  6. Maybe my initial comment was too glib. I grew up in/near Milwaukee, and I thought the grid system was terrific. It was easy to navigate, the abundance of nearby, parallel streets allowed for alternate routings (if one street was blocked, or if you wanted to ride your bike on a calmer street, etc), and the roads connected you to your surroundings (instead of isolating you from them).

    Compare that to, say the Main Line. Only three roads get you anywhere: Lancaster, Conestoga/County Line/Haverford, and Montgomery. Good luck trying to bike or walk anywhere. I appreciate that terrain plays a role, but still.

    A little while ago I lived in Raleigh (no grid) and outside of the city center it was impossible to navigate at night unless you knew your route cold -- random streets at irregular spacings and strange angles. None of them went through, so if you missed a turn you couldn't just take the next one. Dreadful.

  7. Steve -- I really like that urban growth concept. A Google maps planning experiment, centered on some small rural town with existing pikes and country roads, might really flesh out the paradigm in terms of building a city around the existing roads and terrain, rather than the "clean slate" grid approach.